This antitrust case is before us for the second time. See Parts and Electric Motors, Inc. v. Sterling Electric, Inc., 826 F.2d 712 (7th Cir.1987) (hereinafter P & E I). In our view, it should not be. Here is why.
I.
As set forth in our first opinion, when A.O. Smith Corporation acquired defendant-appellant Sterling Electric, Inc. (Sterling) in May 1980, plaintiff-appellee Parts and Electric Motors, Inc. (P & E) was one of the largest of over four hundred nonexclusive wholesale distributors of Sterling electric motors and replacement parts, although it always carried and continues to carry a number of other brands of electric motors and/or parts. At that time, Sterling implemented a new distribution program, under which Sterling contracted with ten “stocking” parts distributors, including P & E, and a number of “referral” parts distributors. By agreement, stocking distributors were able to purchase Sterling parts at lower prices than referral distributors in return for purchasing and promoting minimum quantities of Sterling electric motors. Sterling thus conditioned its stocking distributors’ rights to purchase Sterling parts upon their purchase of a quota of Sterling motors.
In the summer of .1982, Sterling warned P & E to buy more motors or risk termination of its right to buy both Sterling motors and parts. In response, P & E began steering customers toward Sterling motors even though, absent Sterling’s warning, it would have recommended other brands. Although P & E boosted its purchases of Sterling motors somewhat in June and August of 1982, Sterling terminated P & E the following October because it regarded P & E’s level of Sterling motor purchases as insufficient.
On April 4, 1983, P & E filed this action, charging that Sterling had violated section 1 of the Sherman Act, 15 U.S.C. § 1, and section 3 of the Clayton Act, 15 U.S.C. § 14, by tying sales of its electric motors (the “tied” product) to sales of its replacement parts (the “tying” product). After a trial in which P & E presented evidence with respect to the elements of a per se as well as a rule-of-reason tying claim, the jury returned a verdict for P & E. Accord*230ing to its answers to special interrogatories, the jury found that P & E had proved all the elements of its per se claim and, in addition, that Sterling’s conduct amounted to an unreasonable restraint of trade. The jury fixed P & E’s damages at $1,231,992, which amount was trebled to $3,695,976.
Sterling filed an alternative motion for a judgment notwithstanding the verdict or for a new trial, arguing, inter alia, that “even if the Sterling distribution program were a tying arrangement, plaintiff has failed to demonstrate the requisite anticom-petitive impact on the relevant market.” On April 14, 1986, the district court set aside the jury’s verdict for P & E, granted Sterling’s motion for judgment n.o.v., and granted conditionally Sterling’s motion for a new trial contingent upon reversal of the judgment n.o.v. The district court held that the case “should not have been submitted to the jury because, among other things, the evidence clearly shows that P & E failed to satisfy the threshold criterion ‘that the tying seller [Sterling] will acquire market power in the tied product [electric motor] market.’ ” In doing so, the district court relied upon Will v. Comprehensive Accounting Corp., 776 F.2d 665, 674 (7th Cir.1985) (quoting Carl Sandburg Village Condominium Ass’n No. 1 v. First Condominium Dev. Co., 758 F.2d 203, 210 (7th Cir.1985)), cert. denied, 475 U.S. 1129, 106 S.Ct. 1659, 90 L.Ed.2d 201 (1986), an opinion we issued after the jury returned its verdict and judgment was entered in this case, and after Sterling filed its alternative motion for judgment n.o.v. or for a new trial.
P & E appealed from the district court’s ruling, which brings us to our opinion in P & E I, in which the panel held initially that “even if a substantial threat of market power in the tied product is an essential element of a tying case, this point was not raised and preserved in the trial court for purposes of the judgment n.o.v.” Id. at 717. The same was true with respect to the conditional grant of a new trial, the panel held, noting that “the district court conditionally granted the new trial based on a legal theory that Sterling had failed to raise at trial” and that “[e]ven if Sterling should be correct in its current view of the tied market power issue, it cannot obtain a new trial based on a legal issue that it failed to raise at trial.” Id. The panel refused to take the extraordinary action of ordering a new trial on grounds not properly preserved because the “district court’s failure to instruct the jury that market power in the new motor market was an essential element of the claim against Sterling was not plain error,” and because it was not unfair to require Sterling “to live with the theory it presented at trial.” Id. at 718. To the extent the district court’s conditional grant of the new trial relied on the tied market power argument, therefore, it had to be reversed. Id.
The P & E I panel next considered Sterling’s alternative arguments in support of the judgment n.o.v. and the conditional grant of a new trial. Id. Two of Sterling’s alternative contentions are relevant here: (1) that there was no proof of market power in the tying product market; and (2) that there was no proof of any foreclosure of competition in the tied product market. Id. The panel first considered and rejected these alternative contentions as potential support for the judgment n.o.v. It then concluded that, under the instructions given, these same arguments did not provide alternative grounds for a new trial, and fashioned its mandate to the district court as follows:
In this case ... the district court has not stated that the jury’s verdict was contrary to the weight of the evidence when assessed against the instructions that guided the jury’s deliberations. Moreover, the record, viewed from our comparatively remote vantage point, reveals no such imbalance. Nevertheless, we are reluctant to ignore the possibility that the district court perceived such an imbalance but declined to discuss it in the belief that Sterling’s tied market power argument obviated the need to elaborate on these “other things.” We will therefore remand the retrial order for clarification on this point. If the conditional retrial order rested solely on the stated grounds, judgment will be entered for P *231& E. If, on the other hand, the district court also believed, based on some previously unstated but substantial and defensible grounds, that the manifest weight of the evidence favored Sterling when assessed against the instructions under which the jury deliberated, then the court may proceed with the new trial.
Id. at 722.
The district court on remand provided the “clarification” the P & E I panel requested. After “review[ing] the record and the positions presented by the parties,” the district court held that “the jury’s verdict was not against the manifest weight of the evidence when assessed in the light of the instructions that guided the jury’s deliberations.” Consistent with the panel’s opinion in P & E I, the district court entered judgment for P & E.
II.
This is Sterling’s appeal from the district court’s denial of a new trial on remand. Sterling mounts a two-pronged attack. First, Sterling asserts, the district court erred in denying a new trial because “P & E failed to prove that Sterling possesses market power in the tying product market.” Second, Sterling argues, “P & E failed to prove that Sterling forced purchasers to buy a second unwanted product and thereby foreclosed competition.” Sound familiar? We think so. The former assertion resembles a great deal Sterling’s alternative argument for post-trial relief in P & El that “there was no proof of market power in the tying product market.” P & E I, 826 F.2d at 719. The latter strongly resembles Sterling’s alternative argument in P & El that “there was no proof of any foreclosure of competition in the tied product market.” Id.
Ordinarily, matters decided on a prior appeal become the law of the case to be followed on a later appeal. Walsh v. Mellas, 837 F.2d 789, 796 (7th Cir.), cert. denied, — U.S.-, 108 S.Ct. 2832, 100 L.Ed.2d 933 (1988). As the Fifth Circuit has explained, the law of the case doctrine
is based on the salutory and sound public policy that litigation should come to an end. It is predicated on the premise that “there would be no end to a suit if every obstinate litigant could, by repeated appeals, compel a court to listen to criticisms on their opinions or speculate of chances from changes in its members,” and that it would be impossible for an appellate court to perform its duties satisfactorily and efficiently and expeditiously “if a question, once considered and decided by it were to be litigated anew in the same case upon any and every subsequent appeal” thereof.
Id. (quoting from White v. Murtha, 377 F.2d 428, 431 (5th Cir.1967)). Thus, although it is not an “immutable concept,” id. at 789, the doctrine should be applied unless unusual circumstances or a compelling reason render it inapplicable. Chicago & North Western Transportation Co. v. United States, 574 F.2d 926, 930 (7th Cir.1978); see also Shakman v. Dunne, 829 F.2d 1387, 1393 (7th Cir.1987) (“law of the case doctrine ‘is not to be lightly disregarded’ ”), cert. denied, — U.S.-, 108 S.Ct. 1026, 98 L.Ed.2d 991 (1988). These circumstances include “(1) substantial new evidence introduced after the first review, (2) a decision of the Supreme Court after the first review that is inconsistent with the decision on that review, and (3) a conviction on the part of the second reviewing court that the decision of the first was clearly erroneous.” Chicago & North Western, 574 F.2d at 930.
It is, of course, critical to determine what issues were actually decided on the prior appeal in order to define the law of the case. Gertz v. Robert Welch, Inc., 680 F.2d 527, 533 (7th Cir.1982). As a general rule, the doctrine does not extend to issues not presented or decided on the prior appeal and does not include questions present in a case and which might have been decided but were not. Conway v. Chemical Leaman Tank Lines, 644 F.2d 1059, 1062 (5th Cir.1981) (quoting Lehrman v. Gulf Oil Corp., 500 F.2d 659, 663 (5th Cir.1974)); Gertz, 680 F.2d at 533 (“observations, commentary, or mere dicta touching upon issues not formally before the court do not constitute binding determinations”). On *232the other hand, “ ‘if the issues were decided either expressly or by necessary implication, those determinations of law will be binding on remand and on a subsequent appeal.’ ” Conway, 644 F.2d at 1062 (quoting Lehrman, 500 F.2d at 663).
Sterling contends that “this appeal presents different issues in the new trial context than could be developed in the prior appeal.” One “fundamental” difference, Sterling asserts, is in the nature and level of evidence needed to withstand Sterling’s arguments for a new trial as opposed to a judgment n.o.v. Sterling correctly points out that “attenuated” evidence that would preclude granting a judgment n.o.v. can still be insufficient to support a jury verdict under the analysis applicable to a new trial motion.
True. But the P & E I panel was well aware of this. Indeed, it discussed Sterling’s alternative arguments for post-trial relief separately under the different standards of review for judgments n.o.v. and grants of new trials. Notably, before the panel considered Sterling’s alternative contentions in support of the judgment n.o.v., it commented that “the standard of review presented Sterling with a difficult obstacle” because “there must be such a failure of evidence that ‘reasonable men in a fair and impartial exercise of their judgment’ could not have reached the same conclusion as the jury.” P & El, 826 F.2d at 718. And before turning to the issue whether Sterling presented alternative support for the district court’s conditional grant of the new trial, the panel recognized that this court ordinarily “accords great deference to a district court’s grant or denial of a new trial motion,” especially with respect to decisions based on the trial court’s assessment of the weight of the evidence. Id. at 721.
That this appeal now comes before us only in “the new trial context,” then, means nothing. The P & El panel already ruled within both the judgment n.o.v. and the new trial contexts. The problem thus remains that if the issues raised here are in fact the same as those discussed and decided in P & E I, this appeal is precluded by the law of the case doctrine, unless an exception to the doctrine applies.
As noted, the first of the two issues raised by Sterling on this appeal is that “P & E failed to prove that Sterling possesses market power in the tying product market.” The P & El panel had this to say on that point:
[O]n the issue of market power in the tying product, Sterling conceded at trial that it has 100% dominance in the market for Sterling parts and that Sterling parts are unique.' When the seller’s share of the market is high, or when the seller offers a unique product that competitors are not able to offer, the likelihood is that market power exists. [Jefferson Parish Hosp. Dist. No. 2 v. Hyde,] 466 U.S. [2,] 17 [104 S.Ct. 1551, 1560, 80 L.Ed.2d 2 (1984) ].
Id. In a footnote to this passage, the panel elaborated:
The Supreme Court’s characterization of “market power,” see Hyde, 466 U.S. at 13-14, 104 S.Ct. 1558-59, has been alternatively stated in this circuit as “power over price,” or the ability to induce buyers to pay more money by cutting back the supply of goods available for purchase. Will, 776 F.2d at 671. The best way to show power over price is to establish directly that the price of the tied package is higher than the price of components sold in competitive markets. Id. at 671-72. However, in the absence of direct evidence to this effect, a plaintiff may demonstrate market share or uniqueness as proxies for power over price. Id. at 673; see also Times-Picayune Publishing Co. v. United States, 345 U.S. 594, 611-13, 73 S.Ct. 872, 881-83, 97 L.Ed. 1277 (1953) (on market dominance); Fortner [Enters. v. United States Steel Corp.], 394 U.S. [495] at 504-05, 89 S.Ct. [1252] at 1259-60 [22 L.Ed.2d 495] (on tying).
Id. n. 7. Moreover with respect to the tying product market, the panel found that
[additional evidence substantiated Sterling’s market power. In addition to testimony from P & E’s expert witness directly on point, there was evidence that no *233alternative sources existed for Sterling parts; that no one but Sterling manufactured parts compatible with Sterling motors; and that there are effective economic barriers to entry into the market for parts compatible with Sterling motors since the demand for Sterling parts will accommodate only one manufacturer. In addition, P & E adduced evidence that showed that there was general acceptance of the tie among Sterling distributors. The existence of several tying arrangements provides significant evidence of a defendant’s power to restrain free competition.
Id. Finally, the panel also noted that the jury specifically determined in its answer to a special interrogatory that “Sterling ‘possessed the ability to use dominance over the supply of Sterling replacement parts or [sic] force others to buy ... other products,’ ” id. at 719, and that this amounted to a finding of market power in the tying product market, id.
Despite this, Sterling claims that the tying product market issue on this appeal is “substantively” different. Sterling says it demonstrates on this appeal that uniqueness no longer permits a presumption of market power in the tying product market and that “the physical uniqueness of its parts, as opposed to economic uniqueness, would not support the presumption in any event.” But Sterling does not raise a new issue, it merely shoots more bullets at the same target.
What Sterling really is asking on this appeal is for this panel to reconsider the P & E I panel’s use of Hyde, the principal case upon which it relied for the proposition that “uniqueness” can be a proxy for a more thorough market power analysis. We decline to do so. To begin with, it is not even clear whether Sterling argues that the P & E I panel’s use of uniqueness as a proxy for market power was clearly erroneous when it rendered its decision. Indeed, it seems to us that Sterling could not make such an argument, for although high market share alone or the “uniqueness” of a given product may be poor “proxies” for a market power analysis, it cannot be questioned that the majority in Hyde sanctioned the use of these factors to indicate the likelihood of market power and, in turn, to trigger the per se legal approach to tying cases. In any event, under the clearly-erroneous standard, we cannot meddle with a prior decision of this or a lower court simply because we have doubts about its wisdom or think we would have reached a different result. To be clearly erroneous, a decision must strike us as more than just maybe or probably wrong; it must, as one member of this court recently stated during oral argument, strike us as wrong with the force of a five-week-old, unrefrigerated dead fish.
To be clearly erroneous, then, the P & E I panel’s decision must be dead wrong, and we do not believe it is. Although Sterling’s argument that economic, as opposed to physical, uniqueness triggers the Hyde presumption in favor of market power is persuasive, it cannot change our view of this case. First, this argument strikes us as an attack not just on the P & E I panel’s decision, but also the Court’s opinion in Hyde, for it is precisely Sterling’s point— that traits having little economic significance can end up triggering a per se approach to an alleged tie-in — that militates against the use of proxies for market power in tying cases, the practice sanctioned in Hyde. In addition, the panel in P & E I arguably grounded its decision to use uniqueness as a proxy for market power in the tying product on more than just the physical qualities of Sterling replacement parts. The panel did find evidence that other competitors were not able to offer parts compatible with Sterling Motors because of effective barriers to entry in that the demand for Sterling parts could accommodate only one manufacturer. Finally, Sterling contributed to its present dilemma by conceding that its replacement parts were unique. At best, then, Sterling can only argue that we might conclude, after considering how other courts of appeals have construed Hyde and other Supreme Court decisions, that the P & E I panel erred in light of those decisions. But none of the authority cited by Sterling to persuade us to reconsider the panel’s construe*234tion of Hyde is controlling and therefore cannot implicate the “subsequent authority” exception to the law of the case doctrine. On the market power in the tying product issue, then, nothing saves Sterling from that doctrine’s grasp.
The same goes for Sterling’s second argument on this appeal: that “P & E failed to prove that Sterling forced purchasers to buy a second unwanted product and thereby foreclosed competition.” On this, the P & E I panel found
evidence that Sterling’s distributors were forced after the tie-in to increase their purchases of Sterling’s motors by more than $250,000, presumably at the expense of Sterling’s competitors, and testimony that P & E employees engaged in customer “steering,” that is, in diverting customers to Sterling motors over other brands.
Id. at 721. “The effect of Sterling’s tie-in,” said the panel, “could reasonably be perceived by the jury as giving a competitive edge to Sterling products based not on their merits, but on Sterling’s tie.” Id.
Sterling nevertheless argues that the issue is different this time around. According to Sterling, it presented in P & E I a brief “factual” argument that competition in the tied product market was not affected by Sterling’s distribution practices. The thrust of its present argument, says Sterling, is that Sterling’s distribution plan is not “legally or factually of a type that forced P & E to purchase unwanted Sterling Motors.” Again, same issue, new arguments, none of which implicates an exception to the law of the case doctrine. The prior panel’s finding precludes review of any new factual arguments, and Sterling’s new legal argument is not based on any new controlling authority.
In short, this case should not be here. In our view, it represents Sterling’s second attempt to wriggle out of a legal theory it argued unsuccessfully to the jury. We can understand why Sterling has tried to do so. As Judge Posner stated at oral argument, there is no question that this is a threadbare antitrust claim under both the law of tying cases as it is and, certainly, under tying law as certain people argue it should be. But under the instructions given — instructions Sterling must live with — both the district court and the prior panel have now ruled that there was sufficient evidence to support the jury’s verdict against Sterling. And like the panel in P & E I, we do not believe that it is unfair to to require Sterling to live with the theory it presented at trial.
Affirmed.